Trade and greenhouse-gas emissions: How important is international transport?

David Hummels, Anca Cristea, Laura Puzzello 17 December 2011



As the first commitment period of the Kyoto Protocol comes to an end in 2012, member countries of the UN’s Framework Convention on Climate Change (UNFCCC) are meeting in Durban, South Africa, to decide on future actions to curb worldwide greenhouse-gas emissions.

International transport is absent from existing agreements on climate change, and negotiations to include this sector in carbon balances are progressing slowly. Differences in the willingness to regulate the greenhouse-gas emissions from international transport became apparent just a few weeks ago, when air carriers and officials around the world reacted strongly against the EU’s decision to include the aviation sector in its emission-trading scheme (Krukowska 2011).

One of the main difficulties in regulating emissions from international transport is the paucity of data on their magnitude and incidence. The little we know about these emissions comes from the ‘life-cycle analysis’ of very specific products such as Kenyan cut-flower exports. Unfortunately it is difficult to extrapolate from these highly detailed case studies to a systematic evaluation of transport emissions in trade.

In a recent paper (Cristea et al 2011) we provide such an evaluation. The key to our analysis was building a database on how goods move; for every product and country pair we track the share of trade that goes by air, ocean, rail, or truck. This allows us to calculate the transportation services (kg-km of cargo moved), and associated GHG emissions, for every dollar of trade worldwide. Combined with data on GHG emissions from production we can calculate total emissions embodied in exports.

International transport is a significant share of trade-related emissions

In our baseline year of 2004, international freight transport generated 1,205 million tonnes of CO2-equivalent emissions, or 146 grammes of CO2 per dollar of trade. By comparison, production of those traded goods generated 300 grammes per dollar of trade, meaning that international transport is responsible for one third of trade-related emissions.

The aggregate numbers understate the importance of transport for many products. Figure 1 shows the share of transport in trade-related emissions, and it varies significantly over industries. At the low end are bulk products (agriculture, mining), and at the high end are manufactured goods. For important categories such as transport equipment, electronics, and machinery, transport is responsible for over 75% of trade-related emissions. Relatively rapid growth in these industries means that transport emissions will loom ever larger in trade.

Once we include transport, clean producers look dirty

Table 1 provides calculations of output and transport emissions per dollar of trade and shows large differences between regions in emission intensities. Differences in output emissions are driven largely by the commodity composition of trade, with manufacturing-oriented exporters at the low end. Less known and perhaps more surprising are the large differences in transport emissions. The transportation of US exports is nearly eight times more emissions-intensive than the transportation of Chinese exports, and six times more emissions-intensive than Europe.

Table 1. Output and transport emission shares and intensities, by region and country

Note: Total emissions per dollar are calculated as the sum of transport and output emission intensities. *For comparability with transport emissions, output emissions are constructed as a weighted average of sector level output emissions, using trade rather than output weights.

Accounting for transport significantly changes our perspective on which regions have “dirty”, or emissions-intensive trade. India’s production of traded goods generates 143% more emissions per dollar of trade than the US, but after incorporating transportation, its exports are less emissions-intensive in total.

We also see a strong imbalance in transport emission intensities between imports and exports. This is a critical issue for mechanism design when regulating emissions. Do international transport emissions ‘belong’ to the exporter, or to the importer? Given the imbalance shown here, the US would presumably prefer an import-based allocation while East Asian countries would prefer the opposite.

The value of trade is a poor indicator of associated transport emissions

To understand the differences across products and regions shown above, we must recognise that transport emissions depend on the scale and composition of trade. Intuitively, as countries trade more they employ more transportation services and emit more GHG. However, the partner and product composition of trade critically affect the type and quantity of transportation services (kg-km of cargo) employed. When France imports from Japan rather than Germany, a dollar of trade must travel much longer distances. A dollar of steel weighs vastly more than a dollar of microchips, requiring greater fuel (and emissions) to lift. And the choice to use aviation rather than maritime transport involves as much as a factor 100 increase in emissions to move the same cargo. This last fact, along with the unusually large reliance on air cargo in US exports, explains why US exports are so emissions-intensive.

Trade can reduce emissions, in some cases

If two countries have similar emissions from output, then increasing trade (ie shifting from domestic production to imports) will require more international transport and higher emissions. However, if a country with high output emissions reduces production in order to import from a low-emissions country, the savings in output emissions could be enough to offset the higher transport emissions from trade. Which of these cases is most likely? We find that trade flows representing 31% of world trade by value actually are net emission reducers. This happens most commonly in those industries in the left side of Figure 1 – where output emissions are both a large fraction of trade-related emissions and very different across producers. It is much less common in manufactured goods where transport emissions dominate.

Figure 1. The contribution of transport to total trade-related emissions

Eliminating tariff preferences will shift trade toward aviation and maritime transport

With a better understanding of the emissions associated with both output and trade we can examine how changes in trade patterns will affect trade-related emissions over time. In a final exercise we simulated likely trade growth from 2004–20 resulting from tariff liberalisation and GDP growth using a dynamic version of the GTAP model.

The trend toward preferential trade liberalisation in regional trading blocs such as the EU and NAFTA means that tariffs are lower for more proximate trading partners and especially for land-adjacent partners. Rail and truck transport dominates these trade flows. Tariff liberalisation that removes current preferences in favour of a uniform MFN structure will shift trade toward more distant partners (higher kg-km per dollar of trade) and increase the use of aviation and maritime transport. This wouldn’t necessarily raise total emissions (maritime has lower emissions than rail and trucking; aviation much more), but it does mean that a rising share of trade will be outside current monitoring efforts. Getting aviation and maritime transport emissions in the system becomes critical.

Growth in the developing world will cause international transport emissions to skyrocket

We forecast that likely changes due to tariff liberalisation would be somewhat modest but likely GDP growth will yield profound changes in output, trade, and GHG emissions. Our projections have the value of output and trade rising at similar rates, accumulating to 75-80% growth by 2020. International transport services will grow twice as fast, accumulating to 173% growth. Why? Simply put, the fastest growing countries (China, India) are located far from other large markets, and their trade requires greater transportation services.

Some propose that international aviation and maritime transport should be treated as separate entities, essentially countries unto themselves, for purposes of allocating and capping emissions. If this approach is employed, as opposed to including international transport in national allocations or simply taxing the GHG emissions from fuel use, it is difficult to see how future trade growth can be accommodated.

Summary and implications

International transport emissions are a surprisingly large fraction of trade-related emissions that will grow relatively fast as world output increases and trade shifts toward more distant partners. Policymakers must carefully consider how to include international transport emissions in protocols designed to slow emissions growth. Our emission calculations – based on the most accurate trade and transportation data available to date – provide some necessary tools to advance the policy debate.


Cristea, AD, DL Hummels, L Puzzello, M Avetisyan (2011), “Trade and Greenhouse Gas Emissions from International Freight Transport”, mimeo.

Krukowska E (2011), “UN body urges Europe to omit foreign airlines from CO2 curbs”, Bloomberg News, 3 November.

Sim, S, M Barry, R Clift, and S Cowell (2007), “The relative importance of transport in determining an appropriate sustainability strategy for food sourcing. A case study of fresh produce supply chains”, International Journal of Life Cycle Assessment,12:422-431.

Williams, A (2007), “Comparative study of cut roses for the British market produced in Kenya and the Netherlands”, Précis Report for World Flowers, February.



Topics:  International trade

Tags:  Kyoto Protocol, Greenhouse-gas emissions, international transport


   I'm puzzled by the large apparent differences in the carbon intensity of transport and trade. What is the meaning of these numbers? It would make a lot of sense to try and explain the differences. Alternatively, it would make sense to try and compare comparable trades: e.g. export from the US to the EU/China/India and vice versa. It is probably also true that produce with high value per weight (and size?) is more likely to be transported by air. Another important question is where the low-hanging fruit are, when it comes to carbon emissions in transport. Is the impression correct that the US does a lot more transport by air, which if reduced, could make an enormous difference? 
   Does the US in general is running the risk of doing what its car companies did: design gas-guzzling dinosaurs that will eventually make large parts of its industry non-competitive? It must be clear that not putting a price on carbon inside the US doesn't put the US industry at a competitive advantage. It must also be clear that pressure will come to put carbon tariffs on exports from the US, to pay for the CO2 'embedded' in those products. If the tide eventually shifts, the US industry might find itself in a rather awkward position, having to adapt in a panic.
    It seems a quite reasonable thing to start pricing carbon in international transport, especially in shipping and airlines, where fuels have until now not really been taxed at all. This represent a subsidy of these modes of transport as compared to alternatives like truck or train, that are taxed. A clear failure of markets that needs to be fixed. Possibly by imposing carbon tariffs?

Professor of Economics at Purdue University

Anca Cristea

Assistant Professor of Economics, University of Oregon

Laura Puzzello

Lecturer, Department of Economics, Monash University